This repository contains the code for the NVIDIA GTC 2023 Introduction to RAPIDS talk.
NYCTaxi.ipynb contains the code from the talk, and shows how to use the cuDF and cuML libraries in a GPU-accelerated data science pipeline.
NYCTaxi-Blank.ipynb is the same, but contains a few #TODO sections as exercises.
To run the notebooks in this repository, you'll need to install cuDF and cuML along with a few other dependencies.
Below, we show how to use pip to install cudf and cuml. Other installation options can be found here.
pip install cudf-cu11 cuml-cu11 --extra-index-url=https://pypi.nvidia.com
pip install requests seaborn scikit-learn tqdm
pip install hvplot
Don't have access to a GPU, but still want to give RAPIDS a try? You can also run the NYCTaxi notebook on Google Colab.